Gary Benson writes: > Hi all, > > These last few days I've been working on improving aot-compile-rpm. > > The first improvement is the long promised abstraction of the non- > rpm-specific bits: aot-compile-rpm is now joined by a command-line > driven brother, aot-compile, which otherwise does exactly the same > thing. This will hopefully be nice for any of you Debian guys who > want to BC-compile your packages for gcj. > > The second improvement is that instead of invoking gcj itself, aot- > compile and aot-compile-rpm generate makefiles and invoke make on > them. This is of little consequence to aot-compile-rpm, however > aot-compile allows you to pass flags to make, so you can pass it > -j2 and have a faster build. > > ftp://sources.redhat.com/pub/rhug/java-gcj-compat-1.0.53.tar.gz I get some weird results from this: [Loaded (BC-compiled) org.eclipse.core.internal.resources.SaveManager from (file:/usr/share/eclipse/plugins/org.eclipse.core.resources_3.1.2.jar <no certificates>)] [Loaded (bytecode) org.eclipse.core.internal.watson.IElementInfoFlattener from (file:/usr/share/eclipse/plugins/org.eclipse.core.resources_3.1.2.jar <no certificates>)] [Loaded (pre-compiled) org.eclipse.core.internal.resources.IManager from (file:/usr/share/eclipse/plugins/org.eclipse.core.resources_3.1.2.jar <no certificates>)] [Loaded (bytecode) org.eclipse.core.internal.utils.IStringPoolParticipant from (file:/usr/share/eclipse/plugins/org.eclipse.core.resources_3.1.2.jar <no certificates>)] [Loaded (BC-compiled) org.eclipse.core.internal.resources.ProjectDescription from (file:/usr/share/eclipse/plugins/org.eclipse.core.resources_3.1.2.jar <no certificates>)] [Loaded (bytecode) org.eclipse.core.resources.IProjectDescription from (file:/usr/share/eclipse/plugins/org.eclipse.core.resources_3.1.2.jar <no certificates>)] [Loaded (bytecode) org.eclipse.core.internal.utils.Assert from (file:/usr/share/eclipse/plugins/org.eclipse.core.resources_3.1.2.jar <no certificates>)] That some of the classes from the same jar are compiled and some interpreted leads me to suspect that there's something wrong with the way subdivided jars get combined. Andrew.
